SSVEP-Based Answer Selection Using a Low-Cost 8-Channel EEG System

Abstract In this work, the topic of neurogames was explored. The focus was on cost-effectiveness and low complexity. A low-cost 8-channel EEG headset was used for signal acquisition, with only the data from the occipital channels (PO7, Oz, PO8) being utilised. A quiz was developed as the game, in which participants had to select answer options by focusing on visual stimuli flickering at different frequencies. The steady-state visual evoked potential was chosen as the biosignal, which is triggered at the same frequencies as the visual stimuli. Canonical correlation analysis (CCA) was used for decision-making. In this process, the recorded signal data is compared with the two frequencies used for the stimuli. The answer option with the higher coefficient is then selected. Compared to commonly used machine learning algorithms, this method places low demands on the hardware used. Furthermore, it is not only fast to implement but also requires no training data. The experiment was conducted twice: once with the frequency pair 10 Hz and 17 Hz, and once with 6 Hz and 17 Hz. Both trials achieved an accuracy of over 80%.